INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Shiga toxin-producing Escherichia coli (STEC), especially serotype O157:H7, has emerged globally as an important food-borne pathogen. The number of infections caused by STEC has increased significantly since the first reported outbreak in the United States in 1982 (18). The main reservoir of STEC appears to be cattle, but STEC isolates have frequently been found in other ruminants like sheep and goats and sporadically in chickens, pigs, and dogs (5). The dynamics of these pathogens in food-producing animals and the environment is not well understood. Principally, they are transmitted through the consumption of contaminated foods, such as raw or undercooked ground meat products and raw milk (24).

The existing STEC diagnostic assays in Finnish hospitals and in most other countries usually detect only sorbitol and -glucuronidase (PGUA)-negative O157 serogroup. Although strains of serotype O157:H7 are the most important ones in STEC infections, a large number of strains of other serotypes have been implicated in both sporadic cases of infection and outbreaks caused by STEC (24). In some previous studies, non-O157 strains represented 21% (14) and 30% (16) of all human STEC isolates. More than 100 E. coli serotypes have been reported to produce cytotoxins known as Shiga toxin 1 (Stx1) and Shiga toxin 2 (Stx2) (24). In addition to toxin production, another virulence-associated factor expressed by STEC is enterohemolysin (Ehly), a type of E. coli hemolysin that is closely linked to Stx production (4). Most STEC isolates also produce a protein called intimin which is probably responsible for intimate attachment of STEC to the intestinal epithelial cells, leading to the damage of the epithelium. The gene encoding this protein is termed eaeA (1).

The first STEC outbreak in Scandinavia took place in Sweden in the autumn of 1995. More than 100 cases of infection caused by STEC O157:H7 were reported (2). In Finland, the reporting of STEC infections was voluntary until the year 1994. Since then, all laboratories finding STEC have been obliged to send the isolate to the Laboratory of Enteric Pathogens (LEP) of the National Public Health Institute (KTL), Helsinki, Finland.

To evaluate the actual prevalence of STEC strains, including non-O157 strains in patients with bloody diarrhea in Finland, LEP initiated an enhanced nationwide surveillance in 1996. In this paper, we describe the design of our prospective study. In addition, the characteristics of all STEC strains isolated in Finland in the 1990s (to the end of 1997) and studied by several phenotypic and molecular methods are presented.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Collection of strains. Over the 8 years of the study the strains were collected in three consecutive phases, as described below.

(i) Period 1: January 1990 to January 1996, normal surveillance. The diagnosis of STEC infections in Finnish clinical microbiology laboratories (in all 29 laboratories) was based on rather occasional culturing of stools on sorbitol- MacConkey agar (SMAC) and isolation of sorbitol-negative colonies. The pure cultures of these colonies were sent to LEP on a voluntary basis. Since 1994 these laboratories were, however, obliged to send all O157 isolates to LEP on the basis of the requirements established in the Communicable Diseases Act and the regulations of the Ministry of Social Affairs and Health. The strains were stored in skim milk at 70°C until use.

(ii) Period 2: February 1996 to January 1997, enhanced surveillance. On the first of February 1996, LEP started a large-scale investigation of the epidemiology of STEC and other Stx-producing bacteria in Finland. To the clinical microbiology laboratories of 20 acute-care hospitals that cultured stool samples were sent letters offering them free services for the diagnosis of potential STEC infection in all patients with bloody diarrhea. For that, these laboratories were asked to send to LEP by mail the primary SMAC or cystine-lactose electrolyte-deficient (CLED) agar plates with cultures of stool samples from these patients. Examination of the plates for detection of bacteria carrying stx genes encoding Stx1 or Stx2 toxins was undertaken by PCR at least once a week. The cultures were stored at 4°C until they were tested.

(iii) Period 3: February 1997 to December 1997, normal surveillance. Following the enhanced surveillance period, the service that examined stool cultures for bacteria carrying stx genes was offered to all 29 microbiological laboratories in cases in which the patient was hospitalized for bloody diarrhea or suspected STEC infection. The methods used at LEP were the same as those used during period 2.

PCR. The primer sequences selected for the amplification of the stx₁, stx₂, and eaeA genes completely matched the sequences of the corresponding genes encoding Stx toxins and the eaeA gene of EPEC in the GenBank/EMBL database libraries. The oligonucleotides used as primers were purchased from Pharmacia Biotech (Uppsala, Sweden). For PCR, a loopful of growth of gram-negative bacteria taken from the first streaking area of the primary fecal culture plate was suspended in 0.5 ml of sterile distilled water and boiled for 20 min.

Isolation and identification. From the PCR-positive primary fecal culture, distinct E. coli-like colonies and other colonies of gram-negative bacteria were isolated and tested for the presence of the sequence that initially gave a positive result. As many colonies as required (but no more than 100 colonies) to find the isolate carrying these particular genes were assayed. The isolate(s) was subsequently characterized biochemically with the API 20 E system (bioMerieux sa, Marcy l'Etoile, France). Their ability to ferment sorbitol was also tested on SMAC. A test for PGUA was executed according to the manufacturer's instructions (AS Rosco, Taastrup, Denmark).

Serotyping of E. coli. O grouping was carried out by bacterial agglutination (13) with antisera against 35 O groups including O157 and enteropathogenic E. coli O groups (20). Strains giving clumping with 4% saline were defined as rough. The O157 antigen was also tested with the E. coli O157 antigen detection kit (Oxoid, Hampshire, England). Flagellar H antigens were identified by agglutination of motile strains with seven anti-H serum pools which are able to identify all known E. coli flagellar antigens (17) (anti-H sera were kindly provided by Yuli Ratiner from Mechnikov Research Institute for Vaccines and Sera, Moscow, Russia). Strains which were nonmotile after four passages through semisolid agar tubes were defined as nonmotile (HNM). Strains which were O nontypeable with any of the antisera in use at LEP were sent to the International Escherichia and Klebsiella Centre of the World Health Organization at the Statens Seruminstitut (Copenhagen, Denmark) for typing.

Detection of Ehly. The production of Ehly was detected on tryptose agar plates (Difco Laboratories, Detroit, Mich.) supplemented with 10 mM CaCl₂ and defibrinated sheep blood that had been washed three times in phosphate-buffered saline (pH 7.2) (4). The blood agar with unwashed sheep blood was used as a comparison plate. Plates were grown overnight in ambient air at 37°C. A halo around the colonies on the plate with washed blood cells only was interpreted as indicating a positive result for Ehly.

Verotoxin production. The ability of an isolate to produce Stx1 and/or Stx2 was determined by a reversed passive latex agglutination test (Verotox F; Denka Seiken, Tokyo, Japan) according to the manufacturer's instructions; the test enabled detection of toxin titers to a dilution of 1:128.

Antimicrobial susceptibility. Antimicrobial susceptibility of the isolates was studied by the disk diffusion technique on Iso-Sensitest medium by using the zone size criteria recommended by the disk manufacturer and on the basis of the breakpoints established by the Swedish reference group for antibiotics (21). The following antimicrobial agents (AS Rosco, Taastrup, Denmark) were used: ampicillin, ceftriaxone, chloramphenicol, ciprofloxacin, imipenem, mecillinam, nalidixic acid, neomycin, streptomycin, sulfonamide, tetracycline, and trimethoprim.

PFGE. Bacterial cells were grown on Luria agar overnight in ambient air at 37°C. They were suspended in 5 ml of TEN (0.1 M Tris-HCl, 0.15 M NaCl, 0.1 M EDTA) buffer, partially embedded in low-melting-temperature agarose (SeaPlaque agarose; FMC BioProducts), and digested overnight with 10 U of proteinase K (Boehringer Mannheim) at 55°C. The plugs were first washed for 30 min with TE (Tris-EDTA) buffer, then for 60 min with TE buffer plus 200 µl of 100 mM phenylmethylsulfonyl fluoride, and finally, three times for 30 min with TE buffer. Restriction endonuclease digestion was done with 10 U of XbaI (Boehringer Mannheim) overnight at 37°C. Pulsed-field gel electrophoresis (PFGE) was performed with the Bio-Rad GenePath System in a 1% agarose gel in 0.5× TBE (Tris-borate-EDTA) buffer at 14°C with a linear ramp time of 12.6 to 54.20 s over a period of 22 h, a 120° switch angle, and a gradient of 6.0 V per cm. After PFGE, the gels were stained with ethidium bromide and photographed under UV transillumination. An isolate with at least two differences in the banding pattern among bands larger than 100 kb was regarded to belonging to another subtype. Each subtype was given a number within a serotype: serotype O157:H7 isolates were numbered 1a, 1b, etc.; and isolates of serotype O26:H11 were numbered 2a, 2b, etc.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Prevalence of STEC. During period 1, four sporadic infections caused by STEC were diagnosed in Finland (Table 1; Fig. 1). The ages of the patients ranged from 2 to 32 years. Two of these infections were of domestic origin and two were of foreign origin (Spain and Turkey).

View larger version (15K): [in this window] [in a new window]   FIG. 1.   STEC infections in Finland, 1990 to 1997. , non-O157 strains; , O157 strains.

Phenotypic characteristics of isolates. Of all 71 STEC isolates detected in Finland since 1990, 57 were sorbitol and PGUA negative and belonged to the O157:H7 serotype (Table 1). One additional O157 strain was HNM and PGUA positive and fermented sorbitol but did not produce Ehly, in contrast to the other O157 isolates, which produced Ehly. Four strains which were of the O26:H11 serotype also produced Ehly. Of the remaining strains, all (O2:H29, O26:HNM, O91:H21, O91:H40, O107:H27, OX3:H21, and Rough:H49) except two (O101:HNM, O165:H25) did not produce Ehly.

Virulence genes and toxin production. Of the 71 STEC strains, 63 carried the stx₂ gene only, 5 isolates (all were O26) carried the stx₁ gene only, and 3 isolates (all O157) carried both genes. The corresponding toxin(s) shown by the reversed passive latex agglutination test was produced by 69 of 71 strains, with the titers varying from 1:4 to 1:16 for Stx1 and 1:2 to 1:128 for Stx2. The only two isolates which did not produce Stx toxin were O91:H40 and O107:H27. The eaeA gene was detected in all other isolates except five isolates of serotypes O2:H29, O91:H21, O107:H27, OX3:H21, and Rough:H49.

Antimicrobial susceptibility. Of the 12 antimicrobial agents tested, resistance to the following 4 agents was detected: ampicillin, sulfonamide, streptomycin, and tetracycline (either alone or in combinations). Eleven strains possessing resistance markers were of serotype O157:H7 (seven strains) and other serotypes (O26:H11, O26:HNM, O91:H40, and Rough:H49).

PFGE types. All isolates belonging to the serotype with more than one isolate (57 isolates of O157:H7 and 4 isolates of O26:H11) were subtyped by PFGE (Table 1; Fig. 2). Digestion of chromosomal DNA with XbaI yielded 12 to 17 bands. Seven distinct genomic profiles could be seen among the 57 O157:H7 isolates and three distinct genomic profiles could be seen among the 4 O26:H11 isolates. Type 1a was the most common genotype found among 36 (65%) of the O157 strains. All 15 strains isolated during July and August 1997 in western Finland had this type 1a PFGE pattern. Other common O157 subtypes were 1b (seven strains), 1c (six strains), and 1d (five strains). Those subtypes were associated with sporadic cases of infection or cases of infection occurring within a family.

View larger version (112K): [in this window] [in a new window]   FIG. 2.   PFGE patterns of STEC serotypes O157:H7 and O26:H11 isolated in Finland from 1990 to 1997. Lanes 1 and 12, bacteriophage lambda ladder PFG marker 340 (New England Biolabs Inc., Beverly, Mass.); lane 2, subtype 1a; lane 3, subtype 1b; lane 4, subtype 1c; lane 5, subtype 1d; lane 6, subtype 1e; lane 7, subtype 1f; lane 8, subtype 1g; lane 9, subtype 2a; lane 10, subtype 2b; lane 11, subtype 2c.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Since 1994 all laboratories that find STEC in clinical samples have been obliged to report the finding and send the isolate to KTL. However, the systematic reporting of important or interesting microbial findings has been carried out on a voluntary basis since the 1980s (unofficial weekly reports of Orion Diagnostica from 1980 to 1996). The number of identified STEC infections in Finland has been very low. The four sporadic O157:H7 isolates were the only STEC isolates during a 6-year period (1990 to 1995) in Finland, which has a population of 5.1 million. The incidence rate of STEC infection was 0.013/100,000/year; the true incidence, however, was questionable because of the inability of laboratories to detect non-O157 strains. After the first Scandinavian epidemic, in Sweden (2), LEP of KTL started an enhanced microbiological surveillance study of patients with bloody diarrhea to evaluate the actual prevalence of STEC infections in Finland. The methodology used was able to find all STEC types by detection of stx₁ and stx₂ genes. This study confirmed the good situation in Finland: only two sporadic, indigenously acquired O157:H7 infections were discovered during the 12-month study period. Interestingly, however, five of the eight STEC infections that were diagnosed were caused by non-O157 serotypes; of these five infections, three were potentially of domestic origin. After this enhanced surveillance period, for reasons that are yet unknown, the incidence rate for STEC infections changed dramatically within 5 months (incidence rates, 0.06 in 1996 and 0.73 in 1997). In July 1997, the first STEC outbreak in Finland caused by serotype O157:H7 started in the western part of the country (15). A total of 15 microbiologically confirmed cases of infection were found to be associated with this outbreak. These cases still accounted for only 25% of all 61 STEC infections in Finland in 1997. Only one of them could be linked to a certain food item (22). In this case, the patient had drunk unpasteurized milk produced on a neighboring farm. When samples were taken from that farm, the STEC O157:H7 strain isolated from the stool of a cow had a PFGE pattern (subtype 1b) identical to that of the isolate from the patient.

During the whole 8-year study period (from 1990 to 1997), 10 STEC infections were associated with travel outside Finland: Spain and Turkey (2 infections imported from each country), France, Singapore, Russia, and a cruise to Sweden (3 microbiologically confirmed infections).

All 57 O157:H7 isolates found during the study period were negative by sorbitol fermentation and PGUA tests. A distinctive characteristic of the Finnish O157:H7 strains was that they had the eaeA gene and produced Ehly and Stx2 but not Stx1; similar findings for O157:H7 strains have also been described previously among isolates in other European countries. In a Dutch study, 89% of the human STEC O157:H7 isolates were solely Stx2 producing (7). In North America, however, STEC O157:H7 isolates have typically carried both stx₁ and stx₂ (1). Despite the use of molecular diagnostic methods in 1996 and 1997, a single sorbitol- and PGUA-positive O157 strain (O157:HNM) among all 67 STEC strains was detected. The detection of O157 strains with these characteristics is rare globally, probably because tests for the detection of STEC are commonly based on culture on SMAC and sorbitol-positive colonies are not further studied. In a German study, however, when molecular methods were used, as many as 14 of the 104 STEC O157 infections in patients with hemolytic-uremic syndrome were caused by HNM, sorbitol-fermenting, and PGUA-positive strains (6).

The proportion of non-O157 infections among all STEC infections has been notable when diagnostic methods that detect these other STEC types have been used (14, 16). At LEP, the use of the PCR method for the detection of stx genes was begun at the beginning of 1996. This method showed that the proportion (20%) of non-O157 STEC strains among all STEC strains in 1996 and 1997 was noteworthy and analogous to those found in previous studies in Canada (21%) and Belgium (30%) (14, 16).

Of the 14 non-O157 isolates found, serotype O26:H11 was the most common type, with 4 isolates found to be of that type; 1 additional O26 strain was HNM. This serotype has been the most common non-O157 serotype found in some European countries, for example, in Germany and Denmark (24). All O26:H11 isolates found in the present study were Stx1 producing and eaeA, Ehly, and sorbitol positive. In a British study, 70% of 37 STEC O26:H11 strains tested produced Stx1 only and were also Ehly positive (19). Most other STEC strains produce Stx1 very rarely (1); in the current study, no such strain was found. To our knowledge, the other non-O157 STEC strains detected in the present study represented unpublished or unique O:H serotypes (10, 11). These strains were mainly Stx2 producing and sorbitol positive but Ehly negative, and their possession of the eaeA gene varied.

By PFGE, seven distinct genotypes were discovered among STEC O157:H7 strains isolated from 1990 to 1997. Isolates found before 1996 were all of different genotypes. In addition, all O157:H7 strains found during the enhanced surveillance period differed from each other. However, strains of the genotypes of the domestic STEC O157:H7 isolates detected during the enhanced surveillance period have since been found. The 51 STEC O157:H7 strains isolated in 1997 could be divided into four groups of strains according to their genotypes. All STEC O157:H7 strains isolated during the outbreak had similar PFGE patterns (genotype 1a). Strains of the same genotype had already been found sporadically in other parts of Finland 2 months preceding the outbreak. Yet, no connection or common source has been found between the outbreak isolates from western Finland and strains of the same genotype found in different parts of Finland. It seems that a certain STEC clone has become widely distributed throughout Finland. Most interestingly, when the banding pattern of the Finnish outbreak isolate (type 1a) was visually compared to those of the British STEC isolates presented by Willshaw et al. (23), who provided no band size information, identical PFGE patterns were seen. If this is correct, it could mean that this STEC clone was widely distributed in other parts of Europe as well. The number of PFGE genotypes found in this study was lower than the number found, for example, in a recent study in Minnesota (3), where 317 O157:H7 STEC isolates were subtyped by PFGE and XbaI digestion and 143 distinct PFGE patterns were generated with the software that the investigators used. On the other hand, in a recent Japanese study, 825 STEC O157:H7 isolates were similarly subtyped but were classified into only six PFGE types (8). This variation in the numbers of genotypes found in different studies may be due to different interpretation methods, namely, use of a computer versus the eye.

The four O26:H11 isolates could be divided into three different genotypes. To our knowledge, strains of this STEC serotype have not previously been subtyped by PFGE.

In conclusion, during the study period, the incidence of STEC infections in Finland has increased 10-fold. STEC serotype O157:H7 became the major STEC type, as is the case in many other countries in the developed world. Genotyping by PFGE revealed that one type was far more common than the others, but several other types were also found. Moreover, the number of non-O157 infections was notable and constant in both years during which they were investigated. Totally new STEC serotypes were discovered during this study, indicating the need for further studies.